etd@IISc Collection:
http://hdl.handle.net/2005/18
2017-05-26T00:01:56ZMechanism Design For Strategic Crowdsourcing
http://hdl.handle.net/2005/2497
Title: Mechanism Design For Strategic Crowdsourcing
Authors: Nath, Swaprava
Abstract: This thesis looks into the economics of crowdsourcing using game theoretic modeling. The art of aggregating information and expertise from a diverse population has been in practice since a long time. The Internet and the revolution in communication and computational technologies have made this task easier and given birth to a new era of online resource aggregation, which is now popularly referred to as crowdsourcing. Two important features of this aggregation technique are: (a) crowdsourcing is always human driven, hence the participants are rational and intelligent, and they have a payoff function that they aim to maximize, and (b) the participants are connected over a social network which helps to reach out to a large set of individuals. To understand the behavior and the outcome of such a strategic crowd, we need to understand the economics of a crowdsourcing network. In this thesis, we have considered the following three major facets of the strategic crowdsourcing problem.
(i) Elicitation of the true qualities of the crowd workers: As the crowd is often unstructured and unknown to the designer, it is important to ensure if the crowdsourced job is indeed performed at the highest quality, and this requires elicitation of the true qualities which are typically the participants' private information.
(ii) Resource critical task execution ensuring the authenticity of both the information and the identity of the participants: Due to the diverse geographical, cultural, socio-economic reasons, crowdsourcing entails certain manipulations that are unusual in the classical theory. The design has to be robust enough to handle fake identities or incorrect information provided by the crowd while performing crowdsourcing contests.
(iii) Improving the productive output of the crowdsourcing network: As the designer's goal is to maximize a certain measurable output of the crowdsourcing system, an interesting question is how one can design the incentive scheme and/or the network so that the system performs at an optimal level taking into account the strategic nature of the individuals. In the thesis, we design novel mechanisms to solve the problems above using game theoretic modeling. Our investigation helps in understanding certain limits of achievability, and provides design protocols in order to make crowdsourcing more reliable, effective, and productive.2013-12-16T18:30:00ZModel-Checking Infinite-State Systems For Information Flow Security Properties
http://hdl.handle.net/2005/2601
Title: Model-Checking Infinite-State Systems For Information Flow Security Properties
Authors: Raghavendra, K R
Abstract: Information flow properties are away of specifying security properties of systems ,dating back to the work of Goguen and Meseguer in the eighties. In this framework ,a system is modeled as having high-level (or confidential)events as well as low-level (or public) events, and a typical property requires that the high-level events should not “influence ”the occurrence of low-level events. In other words, the sequence of low-level events observed from a system execution should not reveal “too much” information about the high-level events that may have taken place. For example, the trace-based “non-inference” property states that for every trace produced by the system, its projection to low-level events must also be a possible trace of the system. For a system satisfying non-inference, a low-level adversary (who knows the language generated by the system) viewing only the low-level events in any execution cannot infer any in-formation about the occurrence of high-level events in that execution. Other well-known properties include separability, generalized non-interference, non-deducibility of outputs etc. These properties are trace-based. Similarly there is another class of properties based on the structure of the transition system called bisimulation-based information flow properties, defined by Focardiand Gorrieriin1995.
In our thesis we study the problem of model-checking the well-known trace-based and bisimulation-based properties for some popular classes of infinite-state system models. We first consider trace-based properties. We define some language-theoretic operations that help to characterize language-inclusion in terms of satisfaction of these properties. This gives us a reduction of the language inclusion problem for a class of system models, say F, to the model-checking problem for F, whenever F, is effectively closed under these language-theoretic operations. We apply this result to show that the model-checking problem for Petri nets, push down systems and for some properties on deterministic push down systems is undecidable. We also consider the class of visibly pushdown systems and show that their model-checking problem is undecidable in general(for some properties).Then we show that for the restricted class of visibly pushdown systems in which all the high (confidential) event are internal, the model-checking problem becomes decidable. Similarly we show that the problem of model-checking bisimulation-based properties is undecidable for Petrinets, pushdown systems and process algebras.
Next we consider the problem of detecting information leakage in programs. Here the programs are modeled to have low and high inputs and low outputs. The well known definition of“ non-interference” on programs says that in no execution should the low outputs depend on the high inputs. However this definition was shown to be too strong to be used in practice, with a simple(and considered to be safe)“password-checking” program failing it.“Abstract non-interference(ANI)”and its variants were proposed in the literature to generalize or weaken non-interference. We call these definitions qualitative refinements of non-interference. We study the problem of model-checking many classes of finite-data programs(variables taking values from a bounded domain)for these refinements. We give algorithms and show that this problem is in PSPACE for while, EXPTIME for recursive and EXPSPACE for asynchronous finite-data programs.
We finally study different quantitative refinements of non-interference pro-posed in the literature. We first characterize these measures in terms of pre images. These characterizations potentially help designing analysis computing over and under approximations for these measures. Then we investigate the applicability of these measures on standard cryptographic functions.2017-02-15T18:30:00ZTiling Stencil Computations To Maximize Parallelism
http://hdl.handle.net/2005/2619
Title: Tiling Stencil Computations To Maximize Parallelism
Authors: Bandishti, Vinayaka Prakasha
Abstract: Stencil computations are iterative kernels often used to simulate the change in a discretized spatial domain overtime (e.g., computational fluid dynamics) or to solve for unknowns in a discretized space by converging to a steady state (i.e., partial differential equations).They are commonly found in many scientific and engineering applications. Most stencil computations allow tile-wise concurrent start ,i.e., there exists a face of the iteration space and a set of tiling hyper planes such that all tiles along that face can be started concurrently. This provides load balance and maximizes parallelism.
Loop tiling is a key transformation used to exploit both data locality and parallelism from stencils simultaneously. Numerous works exist that target improving locality, controlling frequency of synchronization, and volume of communication wherever applicable. But, concurrent start-up of tiles that evidently translates into perfect load balance and often reduction in frequency of synchronization is completely ignored. Existing automatic tiling frameworks often choose hyperplanes that lead to pipelined start-up and load imbalance. We address this issue with a new tiling technique that ensures concurrent start-up as well as perfect load balance whenever possible. We ﬁrst provide necessary and sufficient conditions on tiling hyperplanes to enable concurrent start for programs with affine data accesses. We then discuss an iterative approach to find such hyperplanes.
It is not possible to directly apply automatic tiling techniques to periodic stencils because of the wrap-around dependences in them. To overcome this, we use iteration space folding techniques as a pre-processing stage after which our technique can be applied without any further change.
We have implemented our techniques on top of Pluto-a source-level automatic parallelizer. Experimental evaluation on a 12-core Intel Westmere shows that our code is able to outperform a tuned domain-speciﬁc stencil code generator by 4% to2 x, and previous compiler techniques by a factor of 1.5x to 15x. For the swim benchmark from SPECFP2000, we achieve an .improvement of 5.12 x on a 12-core Intel Westmere and 2.5x on a 16-core AMD Magny-Cours machines, over the auto-parallelizer of Intel C Compiler.2017-05-20T18:30:00ZComputational And Combinatorial Problems On Some Geometric Proximity Graphs
http://hdl.handle.net/2005/2622
Title: Computational And Combinatorial Problems On Some Geometric Proximity Graphs
Authors: Khopkar, Abhijeet
Abstract: In this thesis, we focus on the study of computational and combinatorial problems on various geometric proximity graphs. Delaunay and Gabriel graphs are widely studied geometric proximity structures. These graphs have been extensively studied for their applications in wireless networks. Motivated by the applications in localized wireless routing, relaxed versions of these graphs known as Locally Delaunay Graphs (LDGs) and Locally Gabriel Graphs(LGGs) were proposed.
A geometric graph G=(V,E)is called a Locally Gabriel Graph if for every( u,v) ϵ E the disk with uv as diameter does not contain any neighbor of u or v in G. Thus, two edges (u, v) and(u, w)where u,v,w ϵ V conflict with each other if ∠uwv ≥ or ∠uvw≥π and cannot co-exist in an LGG. We propose another generalization of LGGs called Generalized locally Gabriel Graphs(GLGGs)in the context when certain edges are forbidden in the graph. For a given geometric graph G=(V,E), we define G′=(V,E′) as GLGG if G′is an LGG and E′⊆E. Unlike a Gabriel Graph ,there is no unique LGG or GLGG for a given point set because no edge is necessarily included or excluded. This property allows us to choose an LGG/GLGG that optimizes a parameter of interest in the graph. While Gabriel graphs are planar graphs, there exist LGGs with super linear number of edges. Also, there exist point sets where a Gabriel graph has dilation of Ω(√n)and there exist LGGs on the same point sets with dilation O(1). We study these graphs for various parameters like edge complexity(the maximum number of edges in these graphs),size of an independent set and dilation. We show that computing an edge
maximum GLGG for a given problem instance is NP-hard and also APX-hard. We also show that computing an LGG on a given point set with minimum dilation is NP-hard. Then, we give an algorithm to verify whether a given geometric graph G=(V,E)is an LGG with running time O(ElogV+ V).
We show that any LGG on n vertices has an independent set of size Ω(√nlogn). We show that there exists point sets with n points such that any LGG on it has dilation Ω(√n) that matches with the known upper bound. Then, we study some greedy heuristics to compute LGGs with experimental evaluation. Experimental evaluations for the points on a uniform grid and random point sets suggest that there exist LGGs with super-linear number of edges along with an independent set of near-linear size. Unit distance graphs(UDGs) are well studied geometric graphs. In this graph, an edge exists between two points if and only if the Euclidean distance between the points is unity. UDGs have been studied extensively for various properties most notably for their edge complexity and chromatic number. These graphs have also been studied for various special point sets most notably the case when the points are in convex position. Note that the UDGs form a sub class of the LGGs. UDGs/LGGs on convex point sets have O(nlogn) edges. The best known lower bound on the edge complexity of these graphs is 2n−7 when all the points are in convex position. A bipartite graph is called an ordered bipartite graph when the vertex set in each partition has a total order on its vertices. We introduce a family of ordered bipartite graphs with restrictions on some paths called path restricted ordered bi partite graphs (PRBGs)and show that their study is motivated by LGGs and UDGs on convex point sets. We show that a PRBG can be extracted from the UDGs/LGGs on convex point sets. First, we characterize a special kind of paths in PRBGs called forward paths, then we study some structural properties of these graphs. We show that a PRBG on n vertices has O(nlogn) edges and the bound is tight. It gives an alternate proof of O(nlogn)upper bound for the maximum number of edges in UDGs/LGGs on convex
point sets. We study PRBGs with restrictions to the length of the forward paths and show an improved bound on the edge complexity when the length of the longest forward path is bounded. Then, we study the hierarchical structure amongst these graphs classes. Notably, we show that the class of UDGs on convex point sets is a strict sub class of LGGs on convex point sets.2017-05-23T18:30:00Z